This is a resubmission of an application for an R21 Grant for Alzheimer's Disease Drug Discovery (PAS-06-261). Accumulation of amyloid-[unreadable] (A[unreadable]) peptides in the brain is thought to play a key role in the pathology of Alzheimer's disease (AD). A[unreadable] peptides are generated as byproducts of the amyloid precursor protein (APP) degradation by [unreadable]-secretase-1 (BACE1). Despite intense drug discovery efforts, few small molecule inhibitors have been developed for BACE1, likely because the active site of BACE1 is more 'open' and less hydrophobic than in many other proteases. Thus, alternative approaches to inhibiting BACE1 have been justified, such as small interfering RNA (siRNA) based strategies. The first aim of this proposal is to develop a non-viral/non- vector based, non-toxic modified siRNA suitable for prolonged knockdown of BACE1 activity in vivo. To accomplish this aim, we will modify conventional siRNAs directed against BACE1 mRNA by incorporation of locked nucleic acid (LNA) nucleotides. LNA incorporation into siRNAs increases stability, resistance to nucleases, and increases siRNA potency and strand specificity. Thus, it is predicted that LNA-modified siRNA directed against BACE1 will decrease BACE1 and A[unreadable] expression in the brains of APP overexpressing mice more potently that conventional siRNA. Recent evidence suggests that, in addition to sense mRNA transcripts, humans and mice also synthesize many natural noncoding antisense transcripts to their sense counterparts. Our laboratory has recently demonstrated that perturbation of antisense transcripts can profoundly alter the expression of their sense mRNA counterparts. Intriguingly, bioinformatic and transcriptomic analyses in our laboratory have now identified a highly conserved BACE1 noncoding antisense transcript, raising the possibility that BACE1 expression levels may be modulated not just by siRNA directed against BACE1 mRNA, but also by knockdown of noncoding BACE1 antisense transcript. This hypothesis is supported by compelling cell culture data and preliminary in vivo data in this proposal. The second aim of this proposal will test the hypothesis that knockdown of BACE1 antisense transcripts will induce a concordant decrease in BACE1 itself in the brains of APP overexpressing mice, potentially revealing a fundamentally new mechanism by which BACE1 expression may be modulated. Overall, we seek to define an optimized non-viral in vivo RNAi approach focusing on BACE1, a prime Alzheimer's disease target. [unreadable] [unreadable] [unreadable]